Prioritizing Circularity in the Design and Manufacturing World

A closed-loop system is essential for the sustainability of product lifecycles and workflows.

The linear versus circular approach. (Image courtesy of Autodesk University.)

The linear versus circular approach. (Image courtesy of Autodesk University.)

Companies are increasingly setting goals to reduce the environmental impact of their products and processes. Last month, published a feature on WNDR Alpine—a ski manufacturer that uses innovative microalgae technology to develop high-performance skis. By replacing petroleum-based products with biomaterial early in the design process, WNDR Alpine is driving trends for companies to embrace circularity and anchor sustainability as a core of their business.

Our planet’s finite resources have traditionally been treated as though they’re unlimited, while the world population grows at an exponential rate. For the longest time, the product lifecycle has followed a linear approach where goods are created, used and thrown away—in a straight line to the garbage dump.

Circularity, on the other hand, refers to a closed-loop system where products are designed to minimize waste and stay in use for as long as possible. When a product following a circular model does reach its end-of-life, its components and materials can easily be reused, remanufactured, refurbished, upcycled and recycled, as opposed to ending up in landfills. Implementing circularity goes far beyond simply recycling; it involves generating an industrial strategy around the entire manufacturing process—from product design, to business models and information workflows, to the supply chain and waste management infrastructure that supports the product lifecycle.

One individual with a passion for circularity is Zoe Bezpalko, Autodesk’s sustainability strategy manager for the design and manufacturing industry. had the opportunity to interview Bezpalko, whose role involves overseeing the development and implementation of Autodesk’s next-generation tools for promoting sustainability.

“Most of my work focuses on the future of our technology,” says Bezpalko. “I work with research and product teams to position our tools so that we’re responding to trends in support of the circular economy as a whole, and particularly as it relates to circular design.”

Bezpalko advocates that transitioning to a successful circular model involves a multifaceted approach towards a product’s design, manufacturing and use phases, followed by its end-of-use stage. In the process, complex problems may need to be solved and inefficiencies may arise, requiring the utilization of advanced technologies such as artificial intelligence (AI).

Circularity in Design

“Circularity starts with design,” asserts Bezpalko. “Design is at the core of sustainability in defining the type of material and construction, whether the product can be easily disassembled, or if its components and materials can be recovered. You can’t recycle something that hasn’t been designed to be recycled.”

The ability to consolidate parts is valuable, as it enhances repairability as well as ease of disassembly. For instance, if plastics are glued together, it can be difficult to take them apart after a product reaches end-of-use. The presence of glue additionally lowers the grade of recycled material at the end of the waste stream. However, if a single source of plastic were used where multiple components were snapped together, this would present an entire set of materials that could be transformed into a new product with relative ease.

Another approach is to design modular parts that can be easily transferred to be reused in other objects. This method of standardization is enforced by a furniture maker called 57st. design, which creates bespoke furniture with elements such as interchangeable chair and table legs. Rather than an entire furniture item needing to be scrapped, materials can be used and products can lead multiple lives, minimizing waste.

Generative design is another key strategy where product design supports circularity. The method leverages AI to design products that meet their functional requirements without the human bias associated with how a product is typically perceived. When implementing the design exploration technology, designers and engineers specify parameters such as design goals, materials, manufacturing methods and cost constraints. The AI-assisted software explores all possible permutations and offers a range of design alternatives. Each iteration provides opportunities for the software to test and learn its solutions, until a final design is attained.

“You only use the material you need for a specific design, so you don’t over-engineer your product,” explains Bezpalko. “Generative design offers excellent sustainability benefits.”

Using AI-driven software to generate relevant design options. (Images courtesy of Whill.)

Using AI-driven software to generate relevant design options. (Images courtesy of Whill.)

One prominent illustration of generative design is the AI Chair—the world’s first production chair created through iterative conversation between a designer and AI. The chair was the result of a collaboration between Philippe Starck, Kartell and Autodesk. An algorithm was utilized to ensure that the least amount of raw material was used, and injection molding was designated as a manufacturing constraint. The chair design was reimagined, while continuing to serve its purpose of providing a comfortable place for people to rest their bodies. Structural durability was maintained, and 100 percent recycled material was used from industrial production leftovers.

 The AI Chair by Philippe Starck, Kartell and Autodesk. (Image courtesy of Autodesk.)

The AI Chair by Philippe Starck, Kartell and Autodesk. (Image courtesy of Autodesk.)

Another instance of generative design was showcased by a company called Whill, an Autodesk customer striving to create a wheelchair that maximized battery life while minimizing weight. Autodesk’s AI software allowed designers to input requirements and constraints, subsequently generating hundreds of options that met the design criteria. After exploring the field of possibilities, the engineers finalized a design that was 30 percent lighter than a traditional wheelchair—with considerably longer battery life as a result, due to the fact that less energy needed to be expended for moving the lightweight wheelchair.

The final design of the Whill wheelchair. (Image courtesy of Whill.)

The final design of the Whill wheelchair. (Image courtesy of Whill.)

According to Bezpalko, intelligent design can drive insights throughout the entire product lifecycle, including a better understanding of waste in manufacturing.

“One of the biggest values of generative design is applying additive manufacturing methods to 3D printing, which has the advantage of printing only the material you need—and therefore, reducing waste in production,” describes Bezpalko.

Circularity in Manufacturing

When it comes to manufacturing, key functionality within Autodesk tools can enhance circularity through improved design. For example, Autodesk Moldflow software enables designers to select alternative materials for injection molding from its database of over 10,000 materials. Sustainable materials include biobased plastics such as polylactic acid (PLA) and Arkema’s Rilsan PA11, a high-performance polymer of 100 percent renewable origin.

“Moldflow also has a recyclability indicator embedded into its material database,” adds Bezpalko.

With technology offering a deeper understanding around the value of sustainable materials, the manufacturing industry can evolve and further realize the energy required for producing items.

Another Autodesk innovation incorporates the use of machine learning to teach robots to build physical objects when provided with a virtual model. In the future, these advancements in production could help manufacturers become more modular and adaptive to their demands.

“Our robotics team have developed AI models that basically build your model in CAD,” details Bezpalko. “You send it through a machine, which recognizes and builds the model in real life. This creates a manufacturing process that is more customizable and therefore leads to less waste.”

Autodesk conducted their robotic research using Lego blocks. (Image courtesy of Autodesk University.)

Autodesk conducted their robotic research using Lego blocks. (Image courtesy of Autodesk University.)

Circular manufacturing involves enhanced connection and information flow. One application is nesting, which refers to the process of laying out cutting patterns from flat sheets of raw material in order to minimize waste.

“The machine cuts shapes into a 2D layout,” explains Bezpalko. “It puts all the shapes together as close as possible, so that there are minimal gaps between them that will be thrown away. Then we have this functionality in nesting called remnant saving. If you’ve created waste—let’s say you use only half of your big metal sheet—you can save those remnants with an identification number in your inventory. The next time you want to perform a flat sheet cutting, you can reuse this waste automatically for your new design. By putting data at the center of the lifecycle, you can leverage it to create insights that incentivize more sustainable, circular design.”

Circularity in the Use Phase

The goal of circularity is to ensure that a product stays in use for as long as possible, maintaining its highest value over time. In the use phase, this can be achieved through a proactive approach towards predictive maintenance.

One Autodesk Foundation customer called SweetSense develops IoT technologies for managing water and energy services in remote, off-grid environments. The company creates connected sensors that leverage machine learning and AI for detecting when a water pump is about to break. Teams are alerted in advance to perform maintenance and ensure that no failures occur, thereby increasing the lifetime of the pumps.

(Image courtesy of SweetSense.)

(Image courtesy of SweetSense.)

SweetSense monitors over three million people’s water supply in East Africa, as well as being employed in California for the compliance of the Sustainable Groundwater Management Act (SGMA). When it comes to water distribution in East African nations, SweetSense secures additional resiliency by utilizing estimated rainfall data from NASA to predict the regions’ needs for water infrastructure. The company’s tagline is: “We fix the Internet of Broken Things.”

Leveraging NASA data for rainfall estimates. (Image courtesy of Autodesk University.)

Leveraging NASA data for rainfall estimates. (Image courtesy of Autodesk University.)

Circularity in End-of-Use

Although recycling is a huge component of circularity, it comes with a multitude of challenges. Inefficient recovery systems, high levels of contamination, lack of standardization and inadequate infrastructure are some factors that pose considerable difficulties around the recycling of plastics. For example, traditional recycling techniques render it impossible to mix different types—or even different colors—of plastics. As a result, less than nine percent of plastic is recycled worldwide. Due to the relative ease with which virgin plastic can be created from inexpensive crude oil, approximately $100 billion worth of single-use plastic ultimately ends up in oceans and landfills every year.

(Image courtesy of Autodesk.)

(Image courtesy of Autodesk.)

Even when products make it to recycling facilities, additional issues abound when it comes to sortation.

“If you visit your local waste management facility, the first thing you’ll see is a conveyor belt with people on both sides,” states Bezpalko. “This conveyor goes at 13mph—so it’s very fast. People have to visually recognize what goes where, and it can be pretty dangerous to handle items in these waste streams. Waste management facilities are having a hard time hiring for these roles.”

After products are sorted, the recycling process continues to be expensive.

“We are mostly reliant on mechanical recycling, where we shred and melt materials into a new shape,” says Bezpalko. “That comes with the challenge of time-cycling the material over and over, and the material loses value.”

As mentioned earlier, products are not often designed to be recycled easily. For example, consider a plastic soap bottle with a metal spring.

“Even if every single material in your soap bottle is recyclable, when it gets on this conveyor belt at 13mph, the person is not going to take the bottle to scrub it and remove its little spring,” attests Bezpalko. “It’s going straight to the landfill. We really need to start thinking about recycling from the design phase, and start building standardization in the way we’re making products, picking materials and so forth.”

According to Bezpalko, one of recycling’s main issues is that designers don’t have adequate information about what can be recycled. Waste management companies, in turn, have limited data about incoming waste streams—leaving them unprepared to adapt their infrastructure for recycling.

One Autodesk partner helping to close the loop is AMP Robotics, a pioneer in AI and robotics that sorts waste into its proper recycling streams. AMP Robotics modernizes the recycling infrastructure by employing its AMP Cortex high-speed robotics system, which identifies and categorizes recyclables from mixed material waste streams at a pick rate above 80 items per minute, with higher accuracy and consistency than humans. The AMP Neuron AI platform differentiates objects by color, texture, shape, size, pattern, opacity and even brand label—continuously training itself as it comes across more items.

AMP Robotics uses AI and advanced data analytics to sort recyclables. (Image courtesy of AMP Robotics.)

AMP Robotics uses AI and advanced data analytics to sort recyclables. (Image courtesy of AMP Robotics.)

“AMP Robotics are so successful that big consumer brands like Unilever are sending them their product and packaging designs before they hit the market, so they can help recycling facilities prepare for the new materials and recycling streams coming their way,” reveals Bezpalko. “It creates much more efficient processes in terms of how much you can recycle, and is safer as well. It’s also a great example of better information transfer throughout the product lifecycle—because now you’re connecting what can be recycled from a design perspective, to how it is actually being picked and sorted out at the management facility.”

Another corporation that has partnered with AMP Robotics is consumer beverage giant Keurig Dr Pepper (KDP), which recently converted its K-Cup coffee pods to recyclable polypropylene. The two companies are working together to ensure that K-Cup pods are recognized by AMP’s robotic systems in recycling facilities.

AI can also be leveraged for enabling circularity in the fashion world—an industry notorious for generating enormous amounts of waste. A company called ThredUp is arranging for clothing to be cycled back into use through resale, donations and recycling—driving circularity by keeping millions of clothes out of landfills. ThredUp uses AI provided by to process items using automated visual tagging, with attributions assigned by neckline, pattern, brand, color, fashion edginess, wear and tear, and much more.

(Image courtesy of ThredUp.)

(Image courtesy of ThredUp.)

Unlike humans, ThredUp can handle massive volumes of clothes thanks to its extensive database, which represents over 35,000 brands. The company, which has processed over 100 million items to date, personalizes resale experiences based on users’ browsing behavior along with the implementation of predictive analytics and trending data. Customers who clean out their closets are rewarded with shopping credit on partnering brands’ sites. If items are beyond the point where they can be worn, they are efficiently sent onwards for recycling into other textile forms.

The Future of Sustainable Manufacturing

Bezpalko believes that methods for obtaining raw materials are currently simpler than the complicated processes around reusing materials. The hurdle is to unlock information and access to recycled materials, making them as easy to obtain as ordering new parts out of a catalog. Circularity can only be established when information flows across the entire lifecycle of products, and value chains are cross-pollinated across industries. Technologies like AI and machine learning offer promising solutions for addressing these complexities.

“There is definitely an uptake happening right now that is unprecedented, particularly due to pressure from consumers,” expresses Bezpalko. “I think the future is going to be where circular design is simply good design, and sustainability is embedded into our design and manufacturing processes as a priority. I foresee a future where products, materials and components are indefinitely circulated—so that we essentially have no waste, no pollution. My dream is for a world that is regenerative by design.”